Postpartum blood pressure trends are impacted by race and BMI

Abstract

Objective:

Our objective was to evaluate postpartum blood pressure trends, and time to resolution of hypertension among women with hypertensive disorders of pregnancy, specifically focusing on impact of race and BMI on these trends.

Methods:

We performed a secondary analysis of a randomized trial that utilized a text-message based home blood pressure monitoring system. BPs for this study included both inpatient postpartum BPs as well as home BPs obtained from the text-based program. Women were followed from 12 h of delivery to 16 days postpartum. Outcomes were: (1) postpartum BP trend summaries from a linear mixed-effects regression model and (2) time to resolution of hypertension (defined as ≥ 48 h of BPs < 140/90) depicted using Kaplan Meier survival curves with hazard ratio estimates of association using Cox models.

Results:

Eighty-four women were included, of which 63% were black. Non-black women with a BMI < 35 kg/m² had steady decreases in systolic BP whereas other groups peaked around 6.5 days postpartum. BPs for women in the BMI < 35 group, regardless of race, remained in the normotensive range. Conversely, women with a BMI ≥ 35 had a systolic BP peak into the hypertensive range prior to declining. Diastolic BP peaked at an average of 8.5 days postpartum. Time to resolution of BPs differed by race and BMI groups (p = 0.012). Non-black women with a BMI < 35 had the shortest time to resolution and 81% of these women had resolution of hypertension. Only 49% of black women with a BMI < 35 had resolution of hypertension and approximately 40% of both black and non-black women with BMI ≥ 35 had resolution of hypertension.

Conclusion:

We identified race and BMI to be determinants of postpartum BP trends and hypertension resolution. Further study is needed to determine if race and BMI targeted postpartum hypertension interventions may lead to faster blood pressure recovery and lower maternal morbidity postpartum.

1. Introduction

Hypertensive disorders of pregnancy (HDP) complicate up to 10% of pregnancies [1]. While there is significant research on the evaluation and management of hypertension during pregnancy, studies on postpartum hypertension (HTN) are usually limited by their focus on inpatients in the immediate postpartum period (2–6 days), or patients readmitted due to hypertensive related complications [2]. Postpartum HTN is associated with significant maternal morbidity including risks of cerebrovascular accidents [3,4], seizures [5], congestive heart failure, pulmonary edema, renal failure and even death [2,6]. Furthermore, postpartum HTN accounts for at least one-quarter of all postpartum hospital readmissions in the United States [7].

Readmissions and maternal morbidity secondary to postpartum HTN are attributed, in part, to delayed mobilization of interstitial and extravascular fluid into the intravascular space [8–10]. Persistent or worsening hypertension is thought to occur 3–6 days postpartum, typically when patients are no longer in the hospital, although data are limited regarding the exact time period of peak blood pressures (BP) and time to resolution of hypertension. Currently, our knowledge on postpartum BP trends and HTN resolution is based on three small studies from nearly 30 years ago, in countries with majority Caucasian populations [9,11–12]. The relevance of these studies in contemporary Obstetric populations must be put into question as there is an increased diversity of pregnant women in the United States [13,14].

In order to improve maternal outcomes for women with HDP, it is critical to fully understand the physiology of postpartum BP in a contemporary cohort to allow for better understanding of the potential role of antihypertensive treatment, as well as mechanisms to improve postpartum monitoring and follow-up. Specifically, as it is well known that obese and black women are at a higher risk for developing HDP and are more likely to suffer a related morbidity, we focused on the impact of race and BMI as two confounders of persistent postpartum HTN [15–21]. Therefore, utilizing post-delivery discharge outpatient blood pressure data from a recent randomized trial evaluating a postpartum text-message based hypertension monitoring program, our objectives were to: 1) evaluate BP trends in the postpartum period and 2) evaluate time to resolution of hypertension among women with hypertensive disorders of pregnancy, specifically focusing on impact of race and BMI on these trends.

2. Materials and methods

We performed a secondary analysis of data from the text-based arm of a randomized controlled trial (RCT) comparing usual care (in-person office based blood pressure checks), to a text-message based home blood pressure (BP) monitoring system. The original study included women with hypertensive disorders of pregnancy (HDP) who delivered from August 2016 to January 2017. For the original study, HDP was defined as any woman with gestational hypertension, preeclampsia with or without severe features, chronic hypertension on antihypertensives, as well as super-imposed preeclampsia with or without severe features as diagnosed per ACOG guidelines [21]. Women randomized to the text-message based program (n = 103) were given an Omron BP cuff and instructed on use prior to discharge from the hospital. On each of the 14 days after discharge, patients were sent text reminders to check their blood pressure and text back values. For ease of use, patients were asked to send in two blood pressures, morning and evening, with no specific times requested. Blood pressure management was standardized based on our institution’s postpartum HTN protocol. This protocol includes starting amlodipine as a first line agent and increasing the dose as necessary to achieve blood pressures < 150/100. Second line agents per this protocol include hydrochlorothiazide and enalapril. Severely elevated blood pressures (> 160/110 mmHg) were managed with either in-person visits or initiation or increase of antihypertensive medication over the phone based on clinical assessment. All patients were asked to attend a 6-week postpartum visit, at which time the need for anti-hypertensive medication would be re-evaluated. Approval was obtained from the University of Pennsylvania Institutional Review Board prior to enrollment.

In this secondary analysis, we excluded women with chronic hypertension (CHTN) without superimposed preeclampsia (n = 2) and women who sent in only 1 day of BP readings after discharge from the hospital (n = 17). Inpatient postpartum blood pressures were obtained starting at 12 h postpartum until discharge using the electronic medical record. Blood pressures sent via text message post discharge were collected for up to 14 days through the texting platform and were included in this analysis.

Categories for BMI were defined using WHO recommendations [23] defined as a three level variable with groups of < 30 mg/kg2, 30–34.9 mg/kg², and ≥35 mg/kg² initially. Given similarities in BP trends for women in the first 2 groups we proceeded with BMI categorized as < 35 mg/kg² versus ≥35 mg/kg² for modeling of BP trends. We considered 4 distinct groups, defined by both BMI and race. Given the small number of women in each category it was not possible to test for a BMI by race group interaction. For our first objective, BP trends over time were analyzed using a linear mixed-effects regression model to characterize change in BP trend over time (in days). Black and non-black groups were based on self-reported race. BMI at delivery was utilized. A quadratic time trend was assumed for modeling, and selection of random effects was driven by likelihood ratio testing. A forward selection process was used to evaluate potential confounders. Variables evaluated to be confounders were nulliparity, gestational age, cHTN with superimposed preeclampsia and mode of delivery. Mode of delivery and cHTN with superimposed preeclampsia were retained in the model as they were determined to be confounders. For our second objective, time to resolution of HTN was evaluated using Kaplan Meier survival curves and hazard ratios (HR) from Cox proportional hazard models. A hazard ratio < 1.0 indicates a slower time to resolution. Time to resolution of HTN was defined as maintaining a systolic BP < 140 mm Hg and diastolic BP < 90 mm Hg for 48 h. For this analysis, we selected the group with the largest sample size (Black with BMI < 35) as the reference group to insure stable parameter estimates for each pairwise comparison, and stable estimates of variability. We conducted a sensitivity analysis where the small group in question (with sample size of 5) was excluded. The resulting analyses were unchanged.

3. Results

From August 2016 to January 2017, 103 women were enrolled in the text-based arm of the RCT, with 95 women texting in at least one outpatient blood pressure. After excluding women with only 1 day of outpatient blood pressure (n = 11), 84 women were included in this analysis. Table 1 shows the demographic characteristics of the study population. The majority of the cohort was black (53 women, 63%) and had a BMI < 35 (65 women, 77%). As seen in table 1, black women were younger (p = 0.001), more likely to be on government (Medicaid) insurance (p < 0.001), less likely to be nulliparous (< 0.001), had higher rates of prior preeclampsia in comparison to non-black women (white n = 25, other n = 6) regardless of BMI (p = 0.014). However, women with a BMI ≥ 35 had higher rates of CHTN compared to women with BMI < 35, regardless of race (p = 0.027). Although black women were more often discharged on antihypertensive medications at discharge in comparison to non-black women regardless of BMI, this was not found to be significant (p = 0.432).

Table 1.

Clinical and demographics characteristics of the cohort.

	BMI < 35		BMI ≥ 35
	Black n = 39	Non-Black n = 26	Black n = 14	Non-Black n = 5
Maternal agea	27.0 (5.8)	31.8 (5.3)	27.9 (6.7)	32.2 (6.1)
Government (Medicaid) insurance	28 (72)	3 (12)	13 (93)	0 (0)
Nulliparous	22 (56)	24 (92)	7 (50)	5 (1 0 0)
Singleton	39 (1 0 0)	23 (88)	14 (1 0 0)	5 (1 0 0)
Gestational Age at Delivery (weeks)	37.5 (3.6)	37.7 (2.3)	35.7 (6.7)	38.2 (1.1)
Preterm delivery < 37 weeks	7 (18)	5 (19)	4 (29)	0 (0)
Birthweight (g)a	2957.2 (839.4)	2844.2 (679.8)	2528.1 (901.1)	3355.0 (365.6)
Pre-gestational diabetes	0 (0)	0 (0)	0 (0)	1 (20)
Prior preeclampsia	11 (28)	0 (0)	5 (36)	0 (0)
Cesarean delivery	18 (46)	9 (25)	5 (36)	0 (0)
Type of hypertensive disorder of pregnancy
Gestational hypertension/ mild preeclampsia	25 (64)	15 (58)	6 (43)	3 (60)
Severe preeclampsia	9 (23)	9 (35)	4 (29)	1 (20)
Superimposed preeclampsia	5 (13)	2 (8)	4 (29)	1 (20)
Antihypertensives at discharge	11 (28)	4 (15)	4 (29)	0 (0)

BMI, body mass index (kg/m²).

Data are presented as n (%) unless otherwise indicated.

^aMean (± SD).

3.1. Blood pressure trends

There was an average of 22.8 BP measurements per patient (range 6–35) for a total of 16 days postpartum (0.5–17.5 days). Systolic BP trends differed by race and BMI grouping (p < 0.001). Fig. 1a shows model estimated systolic BP trends over time. Non-black women with a BMI < 35 had a steady decrease in systolic BP, whereas all other groups had a peak in systolic BP at an average of 6.5 days postpartum. Fig. 1a displays the average systolic BP by group. As seen in the Figure, both black and non-black women with a BMI < 35 had similar systolic BP trajectories which remained in the normotensive range throughout the observation period, although the overall BP average was significantly higher for black women compared to non-black women (p < 0.001) after adjusting for CHTN and mode of delivery. Conversely, both black and non-black women with a BMI ≥ 35 had systolic BP peak into the hypertensive range (> 140 mm Hg systolic) prior to declining in the postpartum period. Diastolic BP trends also differed by group (p < 0.001). Diastolic blood pressure (Fig. 1b) peaked at an average of 8.5 days postpartum, 2 days later than for systolic BP. These trends in both systolic and diastolic BP were consistent after adjusting for CHTN and mode of delivery.

Fig. 1.

a: Estimated systolic blood pressure trajectory by race and BMI. b: Estimated diastolic blood pressure trajectory by race and BMI.

3.2. Time to resolution of postpartum hypertension

In our cohort, only 57% of women achieved BP resolution by 16 days postpartum. However, time to resolution differed by race and BMI groups (p = 0.012) (Fig. 2). Non-black women with a BMI < 35 had the shortest time to resolution (median of 8 days), with 81% having resolution of hypertension by 16 days postpartum. In comparison, only 49% of black women with a BMI < 35 had resolution of hypertension by 16 days postpartum. Approximately 40% of both black and non-black women with BMI ≥ 35 had resolution of hypertension in this 16 day time frame. More specifically, on any postpartum day, the chance that a black woman with BMI < 35 achieved resolution was 50% less likely than a non-black woman in the same BMI category (HR 0.51, 95% CI 0.27–0.95). Meanwhile, the chance that a black woman with a BMI ≥ 35 achieved resolution was about 70% lower (HR 0.29, 95% CI 0.12–0.74). There was no significant difference in time to resolution for non-black women with BMI ≥ 35 compared to non-black women < 35.

Fig. 2.

Time to resolution of hypertension by race and BMI.

4. Discussion

In this unique cohort of women with hypertensive disorders of pregnancy undergoing home BP monitoring for 16 days postpartum, we identified race and BMI to be two specific determinants of postpartum blood pressure trends and hypertension resolution. Non-black women with BMI < 35 had a steady decrease in BP postpartum. All other groups had a peak in systolic BP 6–7 days postpartum (or 8–9 days postpartum for diastolic BP). Importantly, for both black and non-black women, their BP peak often reached the hypertensive range well after discharge. Furthermore, 80% of non-black women with BMI < 35 had complete resolution of their blood pressures by 16 days postpartum whereas only 40–50% of women in the other groups achieved BP resolution in the same time period.

The BP trends seen in this contemporary cohort support the findings from two small studies by Walters et al. from 30 years ago that demonstrated a rise in blood pressures by 3–6 days postpartum in both non-hypertensive women and women with HDP [9,11]. Importantly, however, in the current cohort, we found significant differences in postpartum BP trends and resolution by race and BMI. While our study is one of the first to identify these differences in BP trends, differential outcomes by both race and BMI are seen throughout the obstetrical literature with both black women and obese women noted to be at higher risk of hypertensive disorders of pregnancy [16–21], cesarean [24,25], gestational diabetes [24,26], and postpartum hemorrhage [24,27]. Differences in trends in blood pressures by BMI are not surprising and are supported in the literature in multiple other studies in which a higher pre-pregnancy BMI increases the risk of HDP and, notably, increases BP trends throughout all trimesters [28–30]. Theories regarding the effects of BMI on HDP include higher baseline inflammation [31] and endothelial dysfunction [32] in these patients, which then may contribute to these patients’ higher lifetime risks of CVD.

Strengths of this study include the novel approach to obtaining BP postpartum via a text-based program, which led to the ability to obtain multiple outpatient blood pressures per patient after hospital discharge. Additionally, we followed these women for an entire two weeks postpartum, a time period previously understudied. Furthermore, our cohort included a large percentage of black women, a population at significantly high risk for hypertensive disorders and the associated morbidities.

Our study is limited by our small sample size. While we had 84 women in this study, there were very few women in some categories (e.g. non-black with BMI ≥ 35 had n = 5) which limits our ability to detect more significant differences that may, in fact, be present among the groups. Additionally, this study includes women who delivered at an urban academic medical center which may not be generalizable to other populations.

There is a heightened and important focus on maternal mortality and the postpartum period or “fourth trimester” [33]. Text-based messaging of postpartum blood pressures has been proven to improve adherence to ACOG postpartum BP monitoring, as well as enabled for daily monitoring and communication with patients in regards to their BP in the first two weeks postpartum [22]. As we continue to search for a means to decrease maternal morbidity and mortality in the postpartum period, it is critical to better understand the postpartum physiology and determine if specific demographic or clinical characteristics differentially affect this. Our study sheds light on an important aspect of postpartum blood pressure physiology which we previously knew little about. Moreover, we demonstrated that race and BMI may impact this physiology differently. This information is essential for determining appropriate monitoring and follow-up in the postpartum period for women with hypertensive disorders of pregnancy. For example, understanding that blood pressures peak 6–9 days postpartum, we have the potential to miss a critical time period for intervention without alternative methods to monitoring blood pressures (such as text messaging). Furthermore, understanding postpartum blood pressure physiology can aid in determining the need race or BMI specific treatment modalities. Further study is needed to determine if race and BMI targeted postpartum HTN interventions may lead to faster BP recovery and lower maternal morbidity postpartum in these high risk populations.